In this study, we assessed the impact of a physically-based ice nucleation parameterization and improved cloud macrophysics on the Community Atmosphere Model Version 5 (CAM5) simulated Arctic clouds and climate. The new ice nucleation scheme was developed by DeMott et al. (2010), which links the variation of ice number concentration to aerosol properties while the CAM5 default ice nucleation scheme is parameterized only as a function of ice supersaturation (Meyers et al. 1992). The modified cloud macrophysics is based on an assumed Gaussian PDF of sub-grid scale temperature and moisture variations similar to Sommeria and Deardorff (1977). These changes were tested with CAM5 in hindcasts and free-running simulations. Results were compared to both satellite and ground-based observations. 11-year AMIP-type simulations showed that the new ice nucleation scheme led to a significant reduction in simulated IN number concentrations at all latitudes while changes in cloud amount and cloud properties were mainly seen in high latitudes and middle latitude storm tracks. In the Arctic, there was a considerable increase in mid-level clouds and a decrease in low clouds. The smaller IN concentrations resulted in an increase in liquid water path and a decrease in ice water path due to a slow-down of the Bergeron-Findeisen process in mixed-phase clouds. Overall, there was an increase in the optical depth of Arctic clouds, which led to a stronger cloud radiative forcing (net cooling) at the top of the atmosphere. Results from this study indicate the importance of a better representation of ice nucleation process in mixed-phase clouds in climate models. Linking ice nucleation parameterization to aerosol properties also allows climate models to better represent aerosol-cloud coupling and aerosol indirect forcing. CAM5 hindcasts performed following the Cloud-Associated Parameterizations Testbed (CAPT) protocpl indicated that the improved handling of cloud macrophysics improved simulation of cloud properties and super-cooled liquid water amount, however, did not lead to the best overall simulation, particularly when considering the components of the surface energy budget. This demonstrates the need for future work to better understand and reduce compensating errors in cloud parameterizations.